📖 How to read this report · definitions for every metric below · terms throughout the report are clickable and jump here
Markets tracked
Distinct market keys (e.g. Moneyline|India) seen at any point during the capture for this fixture.
Flapped / Flapped (gone & back)
% of tracked markets that disappeared and then reappeared at least once. High % = unstable feed; the same market keeps blinking out and re-emerging.
Disappearances observed / Disappearances
Total count of DISAPPEARED events across all markets (a single market can disappear/reappear many times). Measures volume of churn, not duration.
Median gap / Longest gap
Across all DISAPPEARED → REAPPEARED cycles, the median and worst time-gap between losing a market and getting it back. "Clean" gaps only — windows of laptop/internet downtime are excluded.
Open-market count / Open-market count over time
How many distinct markets koth-prod / OO AI had live at each moment. koth-prod sampled every ~2s via heartbeats; OO AI at the comparison snapshot cadence (~30s). Dips = bulk drops; flat sections = stable catalogue.
Per-market visibility / Visibility
Per-market timeline showing when each market was offered. Coloured bars = market open; gaps between bars = the market was withdrawn.
Markets missing in OO AI / Coverage gap
Specific markets that other sources priced but OO AI never offered once in this match — genuine coverage gaps. Markets OO AI offered then lost are flapping (see §2), not counted as missing.
Top books filling OO AI's gaps
For each missed market, which books priced it most often (number = how many distinct missed markets that book covered). Useful signal for which markets to source from elsewhere.
All sportsbook availability / Coverage
Per-book coverage of the comparison snapshots in this match. Coverage % = fraction of cmp snapshots where the book had any market; Avg / Peak markets = mean and max market count per snapshot.
Median price level
Median decimal-odds across the common basket (markets all 3 of koth-prod, OO AI, ≥2 books price the same line; basket must have ≥3 markets to count). Each line is a 9-point rolling median (~4.5 min window) to dampen single-snapshot spikes from the basket changing. Gap between lines = systematic pricing divergence.
Cricinfo vs OO lag / Event-detection lag
For each ball-event (wicket/boundary), the time OO AI's REST and Stream feeds detected the event minus Cricinfo's detection time. Below 0 = OO saw it first. The "best" column uses whichever OO feed fired faster.
1 · Open-market count over time
Live selection count over the match. koth-prod at each 2s poll;
OO AI at the comparison cadence (~30s). Dips = markets vanishing in bulk. Red dashed = wickets.
The solid black line marks match start — pre-match time (if captured) is compressed into the left
12% so the in-play action gets the room; axis ticks read −Xh pre-match then +Nm from first ball.
2 · Per-market visibility (top 16 flappers)
visible gone (dropped) no data (laptop) wicket
Top flapping markets · koth-prod
Market
Reappears
Disappearances
Longest gap
Price points
no markets
OO AI exposed no markets in the comparison snapshots.
Top flapping markets · OO AI · sampled at the comparison cadence (~30s)
Market
Reappears
Disappearances
Longest gap
Price points
no markets
5 · Cricinfo vs OO event-detection lag
No matched Cricinfo↔OO events in this log (need match_event records from a live match).
3 · Markets missing in OO AI
Specific markets the other side priced (at any point in the match) that OO AI never offered once —
genuine coverage gaps, not flapping. Grouped by market type; expand a row to see exactly which selections /
lines are absent, and which books priced each. (Markets OO AI does offer but that blink out are flapping — see §2, not counted here.)
vs koth-prod · 0 markets / 0 types
none — OO AI priced every market the others did.
vs books (≥2 books) · 0 markets / 0 types
none — OO AI priced every market the others did.
Top books filling OO AI's coverage gaps · books that priced the most distinct markets OO AI never offered
Book
Missed markets it priced
no per-book attribution available (older capture format)
All sportsbook availability · coverage of comparison snapshots in this match · sorted by coverage %
Book
Coverage
Snapshots with markets
Avg markets / snap
Peak markets
no books seen in this match (capture issue)
4 · Median price level — koth-prod vs OO AI vs books
Median decimal odds on the common basket (markets all three price, mainline, ≥3 markets/basket).
Each line is a 9-point rolling median (~4.5 min window) so single-snapshot basket-churn no longer
spikes the chart. The gap between lines is the pricing divergence; OO AI is dashed so it stays visible
where it overlaps the books line.